Comment by tetraodonpuffer | Hacker News Reader

Comment by tetraodonpuffer | Hacker News Reader

Parent Full thread

tetraodonpuffer

that is a huge pet peeve of mine, the whole nickel and diming and paying developers $xxx,xxx a year and expecting them to make do with a single $200 1080p crappy monitor and the cheapest possible pc one can get.

This seems to happen sometimes because of budgeting issues and silos and so on, but it is still penny wise and pound foolish to expect "passionate" folks and not give them the tools to excel, and creates ongoing friction which will definitely not help with attrition/retention.

If I owned my company I would honestly nowadays give every developer an 8-core 128gig workstation or two with 2-3 21:9 35" 1440p screens and 1TB+ of nvme ssd and a $1k+ allowance to get whatever best-of-the-best chair they want to get (say, a steelcase leap or similar)

Even if you spent $10k on each engineer for this (and the above would be less than that) it would be worth it just from the perspective of you want your employees' workspace to get out of the way as much as possible to enable them to be productive and write good quality code.

The output and code quality gains from having a top-of-the-line machine/monitor pay for themselves in spades, but for some reason this never seems to be noticed in any company I have worked for, it can go from "we are a startup, we have no money to waste" all the way to "we are a large company and beholden to our shareholders, we have no money to waste" but the end result is the same, the people who could do with better tools don't get them and gripe (justifiably) about it.

On the other hand, getting your team spoiled like that can easily make a product that only works for people who have top-of-the-line conditions.

e.g: (2nd hand info - might be wrong) There's a recent Batman game that fails to work properly, and in fact, was essentially pulled from the market and re-released almost unchanged once common-enough gamer machines (still not "very common") caught up with the specs needed for the game.

another e.g: Facebook now does 2G Tuesdays[0] in which network speed is (optionally) purposely degraded to 2G speeds, to make sure that the technologically spoiled Facebook engineers and designers get an idea of the service their lesser-equipped users get.

[0] http://www.businessinsider.com/facebook-2g-tuesdays-to-slow-...

That shows a need for proper test equipment (and guidelines/practices that require those tests), not for an handicapping of the primary tools.

Requiring PC game developers to write a game on a old PC makes no more sense than requiring PSP game developers to write the game on a PSP.

Possibly, but that's a different discussion, and one that -- unlike the "primary tools" discussion, cannot be solved by throwing more money at it.

In my opinion and experience, unless you already have proper testing practices, for PC development it is better to make sure your developers have test rigs on their table that are comparable to end user equipment -- and, in most cases, it's better if it is their development machine.

> Requiring PC game developers to write a game on a old PC makes no more sense than requiring PSP game developers to write the game on a PSP.

"Old" is a spectrum. Most people on this thread say a (e.g. channeling Spolsky) your developer should have the best machine money can buy - with 128GB of ram, 8 cores, 1TB SSD. That's not on the "old" / "new" gamer machine spectrum at all. It's on a completely different scale, given that most gamer machines these days (AFAIK) have 16GB of ram quad cores with 256GB, and casual gamers have 8GB with 128GB SSD or even an HDD

I would require PSP developers to test and debug on a real PSP rather than an emulator - but since PC is both the dev and the target environment, it is reasonable in most cases to require the dev PC to be comparable to the deployment PC, and that would essentially guarantee proper function (at some dev cost, compared to the "best machine money can buy", but with significantly reduced "you need to fix this" testing feedback cost, with a likely overall positive effect -- at least from my experience).

Frankly, I think setting up a standard testing VM with limited resources (2-4 cores, 8GB of RAM, small disk) makes much more sense than purposefully reducing programmer productivity.

In either example, it would probably be more useful to have a combination of dedicated human performance testers, and automated benchmarks that flag problems. Forcing developers to use crappy hardware, even part-time, is more of a kludge for when you don't have a healthy testing and feedback process.

Back when I worked for Telecom Gold ( a pre internet email provider) they deliberately locked our x.25 PAD's down to 1200 or 2400 Baud for the same reason.

It only pays for itself a priori if you are profitable. If you are not profitable then burn-rate is an existential threat which no one will take seriously if the founders don't.

I definitely agree that you should give developers everything they need to maximize their productivity, including spending an extra few thousands on their workstation, but it should really be what they need. For the typical startup I'd really question the need for 1TB SSDs, or the fastest CPUs. Focus on big displays, good ergonomics, ample RAM and good software tools, but that's just me.

1TB SSDs are consumer drives nowadays, if you want to spend money you can get 2+ TB already, same deal for CPUs, you can get 128GB/8core with consumer CPUs and consumer motherboards (asus x99 deluxe, intel 5960x, corsair lxp 128 kit), the higher end would be xeons or dual xeons where you can get a lot more cores/ram for a lot more $$$.

I agree that you can hit a point of diminishing returns, but it also depends on what you work on, being able to spin up a few VMs to do long-running tests on without impacting your main development environment can pay for itself very quickly, IMHO if you hire somebody good at current rates and expect them to stick around for 4 years, 10k in hardware is still only a couple % of what you are going to pay them over that time, which is immaterial in the grand scheme of things, profitable or not.

"For the typical startup I'd really question the need for 1TB SSDs" And that is why you fail. Your already second guessing my needs. I have a 1TB SSD because the 500 was filled. Don't guess. Ask and then supply and don't make me buy it myself.

LOL, I did not fail, I am not your manager. That's why I said "it should really be what they need". Don't smugly put words in my mouth. And while we're being busting out hipster-level snark if you're not willing to put the order in yourself then I start to question is a prima donna who came of age in the boom era and is too precious to do the dirty work that is required to help a company succeed? (http://paulgraham.com/schlep.html )

I think with some edits this could come across more effectively. You're right that the compliant isn't related to what you said, but the way you respond doesn't look very professional. Just my 2c.

Acknowledged, but I've got karma to burn and I'm old enough by SV standards to be have earned my crankiness, so I'll probably continue to take perverse satisfaction by going off from time to time.

Seems like a very good plan. The equipment isn't lost if the employee suddenly leaves and if he/she stays it's really not that big of an investment. Whatever makes them happy and efficient. On top of that it's a pretty decent "lock in" move in an age where good developers are somewhat hard to find.

From anecdotal evidence there are a lot of developers who satisfice when it comes to salary/benefits. There's a certain base level (that they consider "fair pay") and once they get that it's not as simple as adding another 20k/year. Additionally, other factors are way more important (no unnecessary hoops, being happy at work, meaningful/interesting work to do). I've had very heated debates over this but I firmly believe that in "brain activities" specifically, just tossing more money at someone is not exactly the optimal HR strategy.

Hygiene factors vs motivation factors. Some stuff will make you happy, some stuff will make you unhappy. They are not on the same scale. Pay is mostly hygiene - not enough, absolute, or more importantly relative will make you unhappy. It won't make you super happy to increase it over enough though. (Yeah yeah, it totally will for YOU, but in reality... Not really. Increases and relativity matter a lot more than what you really are paid, even me and I'm really cynical about it).

Healthcare, same deal - needs enough to not make you unhappy.

What you work on - can actually make you happier.

When we hire people we just say, "What hardware and software do you want/need?" and take care of it.

Things like a computer, monitor and mouse are such a commodity it's just best to get it out of the way. And the most it's ever cost us is about $3,000 for a top of the line, maxed out MacBook Pro and Thunderbolt Display. But that's a Haskell developer making $150k, so that's a no-brainer, and gets you off on the right foot.

My company gave me a Windows laptop with spinning rust inside. It's 2015, you can get big SSDs for $200 now.

Even without the productivity gains, it's just infuriating to try and do work when I have to wait for the computer.

Right, I bought 2 Samsung EVO 850s yesterday, 256GB for $80 each.

I recently persuaded my company to let me write job postings loudly bragging about the fact that we will buy you whatever fancy hardware you want. Our best hire explicitly mentioned in his cover letter that this caused him to apply.

I don't even know what I'd do with that sort of hardware. Heck, I can barely use 2GB of RAM on a good day, of which the sole significant offenders are Gnome 3 and Firefox (approx. 300MB each).

You are one lucky person. I can't work with anything less than 8 GB + SSD or 16 GB with a regular HDD. Maybe its the result of poor habits (leaving 2 dozen tabs open at any time) and my development stack (RoR + Node + MySQL + Redis).

Reply on news.ycombinator.com